Selectivity and Physicochemical Optimization of Repurposed Pyrazolo[1,5- b]pyridazines for the Treatment of Human African Trypanosomiasis

Article abstract of DOI:10.1021/acs.jmedchem.9b01741

From a high-throughput screen of 42 44 known human kinases inhibitors, a pyrazolo[1,5-b]pyridazine scaffold was identified to begin optimization for the treatment of human African trypanosomiasis. Previously reported data for analogous compounds against h

Full text of DOI:10.1021/acs.jmedchem.9b01741

Subscriber access provided by BIU Pharmacie | Faculté de Pharmacie, Université Paris V
Article
Selectivity and Physicochemical Optimization of Repurposed Pyrazolo[1,5-
b]pyridazines for the Treatment of Human African Trypanosomiasis
Westley Tear, SEEMA BAG, Rosario Díaz González, Gloria Ceballos-Perez, Domingo Isaac
Rojas-Barros, Carlos Cordon-Obras, Guiomar Pérez-Moreno, Raquel García-Hernández, Maria
Santos Martínez-Martínez, Luis Miguel Ruiz-Perez, Francisco Gamarro, Dolores Gonzalez-
Pacanowska, Conor R. Caffrey, Lori Ferrins, Pilar Manzano, Miguel Navarro, and Michael P. Pollastri
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01741 • Publication Date (Web): 17 Dec 2019
Downloaded from pubs.acs.org on December 18, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Selectivity and Physicochemical Optimization of Repurposed Pyrazolo[1,5-b]pyridazines for the
Treatment of Human African Trypanosomiasis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
b
b
Westley F. Tear, Seema Bag, Rosario Diaz-Gonzalez, Gloria Ceballos-Pérez, Domingo I. Rojas-Barros,
Carlos Cordon-Obras,b* Guiomar Pérez-Moreno, Raquel García-Hernández, Maria Santos Martinez-
b
b
c
b
b
b
Martinez, Luis Miguel Ruiz-Perez, Francisco Gamarro, Dolores Gonzalez Pacanowska, Conor R.
d
a
c
b
a**
Caffrey, Lori Ferrins, Pilar Manzano, Miguel Navarro, Michael P. Pollastri
aNortheastern University Department of Chemistry & Chemical Biology, 360 Huntington Avenue, Boston,
MA 02115, USA.
bInstituto de Parasitología y Biomedicina “López-Neyra” Consejo Superior de Investigaciones Cientificas
(CSIC), Granada 18016, Spain.
*
current email address: carlos.cordon-obras@pasteur.fr
cTres Cantos Medicines Development Campus, DDW and CIB, GlaxoSmithKline, Tres Cantos, Spain.
d
Center for Discovery and Innovation in Parasitic Diseases, Skaggs School of Pharmacy and
Pharmaceutical Sciences, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093,
USA.
*
*Corresponding author
Keywords. Lead repurposing; human African trypanosomiasis; neglected tropical diseases; GSK-3β;
CDK-2; CDK-4.
Abstract
From a high-throughput screen of 42,444 known human kinases inhibitors, a pyrazolo[1,5-
b]pyridazine scaffold was identified to begin optimization for the treatment of human African
trypanosomiasis. Previously reported data for analogous compounds against human kinases GSK-3β, CDK-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
and CDK-4 were leveraged to try to improve the selectivity of the series, resulting in 23a which showed
selectivity for T. b. brucei over these three human enzymes. In parallel, properties known to influence the
absorption, distribution, metabolism, and excretion (ADME) profile of the series were optimized resulting
in 20g being progressed into an efficacy study in mice. Though 20g showed toxicity in mice it also
demonstrated CNS penetration in a PK study, and significant reduction of parasitemia in four out of the six
mice.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Introduction
Human African trypanosomiasis (HAT), or sleeping sickness as it is commonly referred to, is a
1
parasitic disease caused by two subspecies of Trypanosoma brucei. Despite sharing a life cycle and causing
similar symptoms in patients these two subspecies, , progress at vastly different rates. T. b. gambiense, is
more common (98% of cases) and infection takes an average of three years to run its course (ending in
fatality), while T. b. rhodesiense, is more virulent and infections can cause death within weeks to months
of infection.2-4 There were 1,446 documented cases of HAT in 2017, with an estimated 13 million people
living in areas of moderate to high risk of infection.1-2 Fatal if left untreated, HAT has two stages; in the
first stage T. brucei resides and multiplies in the blood and lymphatic systems. Infections often are not
diagnosed due to nondescript symptoms, commonly associated with the flu.1, 3 In the second stage, the
parasite crosses the blood-brain barrier (BBB) to infect the central nervous system (CNS), and those
infected present more readily diagnosable symptoms such as disruptions to sleeping patterns, cognitive
dysfunction, and can become comatose.1
There are currently two approved therapies for stage 1 HAT, pentamidine for T. b. gambiense
infections and suramin for T. b. rhodesiense infections, both are ineffective against stage 2 of the disease.2
In the past, the two principal treatments for stage 2 of the disease were eflornithine for T. b. gambiense and
melarsoprol for T. b. rhodesiense: the mechanisms and drawbacks of each is discussed in a thorough review
5
by Barret et al. In 2009, a nifurtimox-eflornithine combination therapy (NECT) was approved for treatment
6
of T. b. gambiense, infections. NECT requires 14 IV infusions of eflornithine (400 mg/kg/day) over 7 days
ACS Paragon Plus Environment

Page 3 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
as well as oral nifurtimox 3 times a day for 10 days.1-2, 7-8 Fexinidazole, a nitroaromatic compound which
recently concluded a Phase II/III clinical trial for treatment of T. b. gambiense infections demonstrated a
9
9
1% cure rate after 10 days of treatment, compared to 97% for NECT therapies. This lower cure rate was
deemed acceptable as fexinidazole has a more manageable dosing regimen due to it being orally
bioavailable, thus fexinidazole was approved for distribution in 2019 by the European Medicines Agency,
becoming the first oral therapy for HAT.3, 10-11 Notably, T. b. brucei cell lines which were resistant to
nifurtimox (also a nitroaromatic compound) were found to be resistant to fexinidazole.12 This cross-
resistance could imply similar mechanisms of action, meaning T. b. gambiense strains resistant to NECT
could also be resistant to fexinidazole.12-13 Acoziborole (SCYX-7158), which recently had a successful
Phase I clinical trial could prove to be the first, single-dose, oral therapy for HAT, however, it is not
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
14
expected to conclude Phase II/III trials until 2020. While promising, it should be noted that treatments for
1
5
infectious diseases fail at a higher rate in the clinic than other drug discovery programs. Should resistance
to NECT/fexinidazole emerge, or acoziborole fail in clinical trials, there would be few remaining treatment
options for HAT.
Despite the need for treatment options, there has been little investment from the pharmaceutical
industry due, in-part, to a lack of financial incentive, which led to the designation of HAT as a neglected
tropical disease by the World Health Organization.16-17 With the lack of investment there has been an
increasing effort in academic settings, and in industrial-academic partnerships to aid the drug discovery
18
process. Traditional drug discovery programs can be time consuming and costly, often starting from a
high-throughput screen (HTS), and going through target validation, hit-to-lead optimization and in vivo
testing before advancing to clinical trials.17 We hypothesize that employing one of a variety of drug
repurposing strategies can help to shorten the timelines associated with this process.17
T. b. brucei expresses 176 kinases (156 which are described as eukaryotic protein kinases). Many
of these are crucial to survival of the parasite; and several have human orthologs which have been pursued
in drug discovery.19-22 With this in mind, we undertook a lead repurposing study against T. b. brucei using
1
8
a biased library of known human kinase inhibitors. Starting from a library of 42,444 kinase inhibitors we
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
identified 797 compounds which showed sub-micromolar activity against T. b. brucei, as well as > 100-
fold selectivity over human HepG2 cells (used as a toxicity counter screen). Compounds were clustered
based on structural similarity and scored based on a variety of properties: their activity against T. b. brucei,
rate and reversibility of action in parasite growth inhibition, and predicted CNS penetration as predicted by
the multiparameter optimization (MPO) score; these properties were folded into a composite score to help
prioritize starting points for medicinal chemistry optimization.18, 23 One of these clusters contained a
pyrazolo[1,5-b]pyridazine: a representative compound is shown in Table 1. Its properties are listed,
including those known to influence drug absorption, distribution, metabolism, and excretion (ADME), such
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
as human liver microsome (HLM) and rat hepatocyte (RH) intrinsic clearance (Cl ), thermodynamic
int
aqueous solubility (aq. sol.) and human plasma protein binding (PPB).
Table 1. Representative hit from the T. b. brucei HTS.18
N
N
N
N
N
N
H
CN
1
Targeted
1
Properties
> 7.5
< 5
T. b. brucei pEC50
MRC5 pTC50
HepG2 pTC50
Aq. Sol. (µM)
MW
8.2 ± 0.02
< 4.3 ± 0.00
< 5
< 4.0 ± 0.00
> 10
≤ 360
≤ 3
≤ 2
≥ 4
≤ 9
≤ 5
≤ 95
< 0.1
310
2.8
3.6
5.1
66
cLogP
cLogD
MPO Score
*
RH Clint
*
*
HLM Clint
210
98
PPB (%)
Aq. Sol. = Aqueous solubility.
MW = Molecular weight.
MPO = Multiparameter optimization.
RH = Rat hepatocytes.
HLM = Human liver microsomes.
ACS Paragon Plus Environment

Page 5 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PPB = Plasma protein binding.
6
*
*
= (µL/min/10 cells).
* = (µL/min/mg protein).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Based on the properties of 1, this class of pyrazolo[1,5-b]pyridazines was considered a promising
starting point for treatment of both stage 1 and 2 of HAT according to the published target product profile.24-
25
This scaffold had already been investigated for activity against serine/threonine kinase (Stk1) in
Staphylococcus aureus and for three human kinases: glycogen synthase kinase 3 beta (GSK-3β), and cyclin
dependent kinases (CDK) 2 and 4.26-28 This was unsurprising as the compound screening set used in the
HTS had been selected from a library of known kinase inhibitors, and homologs of both GSK-3 and CDK
have been identified in T. brucei as potential targets for drug discovery.21, 29
With this information, we set out to answer the following questions. First, would we be able to
discern a difference between the series activity against T. b. brucei cells and human enzymes GSK-3β,
CDK-2 and CDK-4 by utilizing available data for compounds in this class? Second, would understanding
this difference enable us to design compounds with improved activity against T. b. brucei and reduced
potency against these human kinases? Third, would compounds from this class demonstrate efficacy in
mouse models of trypanosome infection? While we used known human kinase inhibitor data to help guide
our molecular design, we did not infer anything about the parasitic target from this information, as previous
work optimizing for activity against T. brucei GSK-3 has demonstrated anti-parasitic activity to have arisen
from a multi-target effect.30
Results
Selectivity
In order to assess the difference in compound potency between T. b. brucei and GSK-3β, CDK-2
and CDK-4, we evaluated previously reported compounds from the original HTS (and synthesized a set of
previously reported analogs) with potency data against the three human kinases, to see if we could find any
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
difference in the selectivity profiles.26, 31 A synthetic strategy was devised (Scheme 1) that would allow for
1
3
late stage derivatization at the three positions of principal interest (R -R , as shown in 8). Starting from 2,4-
dichloropyrimidine (5), a Sonogashira coupling was employed with trimethylsilyl acetylene. Subsequent
silyl deprotection with potassium hydroxide yielded 2-chloro-4-ethynylpyrimidine 6b. Pyridazine 4 was
reacted with hydroxylamine-O-sulfonic acid (HOSA) to generate a 1-aminopyridizinium ion in situ, which
was then reacted with 6b in a [3+2] cycloaddition reaction to form the 3-(2-chloropyrimidin-4-
yl)pyrazolo[1,5-b]pyridazine intermediate 7. Subsequent nucleophilic aromatic substitution or palladium-
mediated cross-coupling provided the corresponding product 8.32
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthesis of pyrazolo[1,5-b]pyridazine analogs.
H
N
O
N
Cl
N
R2
a
b
N
N
N
2
3
4
R2
R2
R1
N
N
N N
N N
R1
R1
Cl
c
N
e
f or g
N
N
N
R3
N
Cl
N
Cl
N
Cl
N
8
N
1
R =TMS 6a
H
5
d
7
1
R =H 6b
Reagents and conditions: a) POCl , 80 °C, 1 h, 70%; b) NaOtBu, MeOH or morpholine, n-BuOH, 80 °C,
3
1
12 h, 60-100%; c) HCCR , CuI, Pd(PPh ) Cl , TEA, THF, 50 °C, 12 h, 19-61%; d) KOH, MeOH, RT, 30
3
2
2
min, 94%; e) 4, hydroxylamine-O-sulfonic acid, KOH, NaHCO , H O, DCM, 70 °C→RT, 12 h, 40-84%;
3
2
3
3
f) NH R , n-BuOH, 150 °C, 2 h, 11-90%; g) NH R , Pd(OAc) , BINAP or Xantphos, Cs CO , dioxane,
2
2
2
2
3
1
60 °C, 1 h, 6-70%.
Figure 1. Previously reported crystal structure of pyrazolo[1,5-b]pyridazine compound, (R)-1-
(dimethylamino)-3-(4-((4-(6-morpholinopyrazolo[1,5-b]pyridazin-3-yl)pyrimidin-2-
ACS Paragon Plus Environment

Page 7 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
yl)amino)phenoxy)propan-2-ol, 9 (green, stick representation), bound in human CDK-2 (grey, surface/wire
31
representation) (PDB: 3EID) displaying; (a) structure of 9, (b) binding mode of 9 within the binding
pocket, highlighting key hinge binding interactions with LEU-83, (c) binding position of 9 showing ligand-
1
surface interactions within 5Å, displaying position of R inside the binding pocket.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(a)
O
N
N
N
N
N
OH
O
N
N
N
H
9
(b)
(c)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A previously reported crystal structure of a pyrazolo[1,5-b]pyridazine, 9 (Figure 1a), bound into
1
2
CDK-2 (PDB: 3EID) (Figure 1b and 1c), indicated that substitution at the R and R positions gave the
best chance to improve selectivity, due to the increased interaction with the binding pocket at these
3
positions, while the R position was solvent exposed, and the binding pocket would be less sterically
31
restrictive of the molecule. Previously-reported docking of a pyrazolo[1,5-b]pyridazine into a GSK-3β
crystal structure (PDB: 1GNG) predicted a similar binding mode.26, 31 Available IC data against all three
5
0
enzymes indicated that substitution at any of these three positions could cause significant shifts in potency,
1
2
but the largest decreases in potency came with substitution at the R and R positions. Armed with the
published human kinase selectivity data, we compared compounds from the HTS and resynthesized a
selected set of analogs from the literature. We sought to identify which substitutions could provide analogs
with divergent activity for T. b. brucei over GSK-3β, CDK-2 and CDK-4 (Table 2).
Table 2. Pyrazolo[1,5-b]pyridazine analogs with published human kinase data, compared with whole cell
T. b. brucei potency data.
R2
N
N N
R1
N
R3
N
N
H
ACS Paragon Plus Environment

Page 9 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
T. b. brucei GSK-3β CDK-2 CDK-4
ID
R1
R2
R3
*
pEC50
pIC50
pIC50
pIC50
10a
-H
-H
5.2
5.031
6.031
6.031
Cl
N
1
0b
0c
-H
-H
-H
-H
4.8
6.4
4.531
--
5.531
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5.831
6.931
7.131
1
O
4.631
6.431
1
0d
0e
0f
-H
-H
6.4
7.4
5.3
--
N
1
-H
7.726
8.326
6.826
--
CF3
1
-H
-H
-H
-H
< 4.526
< 4.726
1
0g
1
-H
-H
-H
-H
-H
5.7
8.2
8.2
7.6
5.7
5.226
8.031
9.531
8.026
7.326
--
--
8.731
9.531
8.726
6.726
8.431
8.731
7.226
< 4.726
CN
1
0h
NO2
1
0i
-H
O
O
1
0j
CF3
O
O
1
0k
-H
-H
-H
7.2
7.926
7.826
7.726
5.826
O
1
0l
-H
-H
6.1
9.026
8.526
F
F
8.026
4.626
9.026
8.726
< 4.726
8.226
10m
0n
10o
-H
-H
-H
6.8
6.1
7.2
--
--
6.526
1
-H
O
O
7.326
< 4.726
1
1
0p
0q
-H
-H
5.8
6.0
--
--
CF3
O
5.326
< 4.726
N
N
1
0r
-H
-H
7.8
7.831
8.431
8.231
CF3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
0s
-H
-H
-H
6.2
7.326
7.426
6.826
5.726
--
10t
9
-CH3
5.6
< 4.726
CF3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OH
O
O
N
6.7
6.231
6.431
7.431
N
-
*
Compounds without published data against a human kinase data were left blank.
standard error of mean (SEM) within 0.17.
1
Substitution at the R position caused varying shifts in potency between T. b. brucei and the human
kinases of interest. In general, substitution at this position led to a significant loss in potency against CDK-2,
2
6
1
an observation that is consistent with the literature. While substituting R with 4-(trifluoromethyl)phenyl
(10f) caused a 100-fold decrease in potency against T. b. brucei (compared with 10e) there was a much
greater impact on the potency against GSK-3β and CDK-2, which both exhibited a > 1,000-fold decrease.
Similarly, substitution with 4-methoxyphenyl (10q) caused a 10-fold decrease in the potency against T. b.
brucei (compare with 10o), however the potency against GSK-3β and CDK-2 was decreased by > 10,000-
fold. Replacement of the phenyl ring with a (smaller) cyclopropyl group (10n) also resulted in a slight
decrease in the potency against T. b. brucei relative to the unsubstituted analog (10m). This loss in potency
was again much less dramatic than observed against GSK-3β and CDK-2 (between 3.5- and 4-log units).
This indicated that while small substituents were tolerated, there was a sharp decrease in activity against T.
1
b. brucei as R became larger. In sum, we were able to demonstrate the potential of compounds with T. b.
brucei activity that is divergent from human GSK-3β and CDK-2 activity particularly when substituted at
1
2
R or R .
2
Substitution at the R position with a morpholine group (10d) led to an equipotent compound
against T. b. brucei relative to the unsubstituted analog (10c), with a concomitant and significant decrease
in activity against both GSK-3β and CDK-4 (~10× against both). The presence of a methoxy at the R2
ACS Paragon Plus Environment

Page 11 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
position (10j) led to a dramatic decrease in activity of the series against T. b. brucei; it also contributed to
a complete loss of activity against CDK-4.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Preliminary structure-activity relationship (SAR) trends for substitution of aromatic rings at R3
position.
Di-substitution with the
3-position not tolerated
Di-substitution with the
3-position tolerated
N
- EWG most active
- EDG have no effect
H
Methylation led to loss
in activity
3
Substitution at R caused a significant shift in the potency of the series against T. b. brucei. Analogs
3
with an alkyl attachment at the R position (10a, 10b, 10c, 10d) showed moderate inhibition of T. b. brucei,
though aryl substituents (summarized in Figure 2) caused a noticeable increase in potency against T. b.
brucei (typified by 10e). While these changes caused significant shifts in the activity of the compounds,
they did not provide any tangible selectivity improvement against the three human kinases. Substitution of
the aryl ring at the 3-position with electron withdrawing groups (EWG) yielded the most potent compounds
against both T. b. brucei and the human enzymes (1, 10h). Electron donating groups (EDG) at the 3-position
(10i) did not cause any significant shift in the potency, though di-substitution at the 3- and 5-positions (10k,
1
0l), caused a noticeable decrease in activity against T. b. brucei. Di-substitution at the 3- and 4-positions
of the phenyl ring (10m, 10r) did not seem to cause this shift in potency, this trend implied there were less
3
steric restraints on the 4-position of the R ring for activity against both T. b. brucei cells and the human
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
enzymes. This observation is consistent with the crystal structures for human GSK-3β, and CDK-2, wherein
the 4-position of the ring points out of the binding pocket.26, 31
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Synthesis of pyrazolo[1,5-a]pyridine analogs.
N N
N N
N
a
b or c
N
N
R3
N
Cl
N
N
H
N
Cl
6b
11
12
Reagents and Conditions. a) pyridine, hydroxylamine-O-sulfonic acid, KOH, NaHCO , H O, DCM, 70
3
2
3
3
°
C→RT, 12 h, 51%; b) NH R , K PO , XPhos, Pd (dba) , toluene, 110 °C, 12 h, 22%; c) NH R , n-BuOH,
2 3 4 2 3 2
110 °C, 12 h, 40-88%.
Curious to the role of the pyrazolo[1,5-b]pyridazine head group to the activity of the cluster, a
series of analogs was prepared to elucidate the impact of the nitrogen atom positioning within the head
group upon T. b. brucei activity. The pyrazolo[1,5-a]pyridine head was synthesized via Sonogashira
reaction using trimethylsilylacetylene (Scheme 2). Following desilylation, a [3+2] dipolar cycloaddition
was effected using a 1-aminopyridinium ion generated in situ, to give 3-(2-chloropyrimidin-4-
yl)pyrazolo[1,5-a]pyridine (11) in moderate yield. Subsequent nucleophilic aromatic substitution gave the
desired products.
Table 3. Effect of pyrazolo[1,5-a]pyridine head on T. b. brucei potency ADME properties.
ACS Paragon Plus Environment

Page 13 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
N N
N
R3
N
N
H
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
T. b. brucei
Aq. Sol.
(µM)
RH
Clint
HLM PPB MRC5
ID
R3
LogD 7.4
*
**
***
pEC50
Clint
(%)
pTC50
1
2a
6.0
6.6
64
2.9
160
75
91
< 4.3
1
2b
2
4
76
140
63
> 300
220
99
> 99
99
< 4.3
< 4.3
< 4.3
1
2c
6.6
7.0
18
4.3
OMe
CN
1
2d
< 0.9
> 3.9
nda
*
*
= SEM within 0.05.
* = (µL/min/10 cells).
*** = (µL/min/mg protein).
nd = not determined.
6
a
compound detected only in first sample.
As Table 3 shows, replacement with the pyrazolo[1,5-a]pyridine head caused a decrease in potency
for all four analogs compared to their matched pyrazolo[1,5-b]pyridazine counterparts. Compound 12a
showed the smallest decrease in potency compared to 10c but was still 2-fold less potent. The aromatic
3
substituents at the R position were a minimum of 7-fold less active against T. b. brucei (12b compared to
1
0e), all the way up to 15-fold less active (12d compared to 1). Due to this decrease in activity, and no-
tangible gain in the ADME profile of the series, we turned our attention to replacing the nitrogen atoms in
the pyrimidine ring.
The crystal structure and docking study for both GSK-3β and CDK-2 indicated that the key hinge
binding interaction (highlighted in Figure 1b) with the series occurred between the pyrimidine nitrogen,
3
26, 31
and amine at the R position.
If altering or removing this interaction did not cause a significant loss in
potency against T. b. brucei, it would likely result in a large selectivity improvement over the human
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
enzymes. As such, we set out to remove or replace the pyrimidine and amine nitrogens. The pyridine
analogs (shown in Table 4) were prepared as shown in Scheme 1, starting from the corresponding
ethynylpyridine. However, this route proved unsuccessful for the phenyl derivatives, so a new synthetic
route was employed, shown in Scheme 3. Starting from pyridazine, a [3+2] dipolar cycloaddition was
performed with methyl propiolate. The subsequent product was saponified using potassium hydroxide to
give the corresponding acid. A Hunsdiecker-like transformation provided the bromide, which was then used
to perform a Suzuki-cross coupling to install the phenyl ring.33
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of pyrimidine replacement analogs containing phenyl ring.
N
N
N N
N N
N
N
N
N
N
N
a, b
c
d
e
CO Me
2
HO C
Br
2
N
H
CN
X
1
6a X = H
13
14
15
16b X = Cl
17
Reagents and conditions. a) pyridazine, hydroxylamine-O-sulfonic acid, KOH, NaHCO , H O, DCM, 70
3
2
°C→RT, 12 h, 71%; b) NaOH, H O, 70 °C, 1 h, 37%; c) NBS, DMF, 25 °C, 3 h, 79%; d) R(BOH) ,
2
2
Pd(dppf)Cl ·DCM, K CO , 3:1 dioxane/H O, 130 °C, 1 h, 15-72%; e) 3-aminobenzonitrile, Pd(OAc) ,
2
2
3
2
2
Xantphos, Cs CO , dioxane, 160 °C, 1 h, 47%.
2
3
Both pyrimidine nitrogen atoms were deemed crucial for potent activity (Table 4), as removal of
either (18d, 18b) resulted in a > 10-fold potency loss compared to the original hit (1). Removal of both
nitrogen atoms (17), resulted in an additional decrease in potency. Removal of the hydrogen bond donor
3
(
HBD) and R ring (16a, 18a, 18c) resulted in a total loss of activity against T. b. brucei, though it also
corresponded with a substantial increase in the solubility of all three compounds.
ACS Paragon Plus Environment

Page 15 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. T. b. brucei activity of the pyrimidine replacement analogs, and the associated ADME data.
N
N N
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Y
X
R3
T. b. brucei Aq. Sol. LogD
RH
Clint
66
HLM PPB
MRC5
ID
6a
R3
X
Y
*
**
***
****
pEC50
(µM)
7.4
Clint
(%) pTC50
1
-H
CH CH
CH CH
< 4.4
850
2.8
> 300
91
99
61
99
86
99
< 4.3
1
7
5.4
< 4.4
6.2
0.7
> 1000
0.6
4.2
1.9
4.1
1.2
3.7
82
57
84
5
300
170
280
37
4.7
N
H
CN
CN
CN
18a
18b
18c
18d
-H
-H
CH
CH
N
N
N
< 4.3
< 4.3
< 4.3
< 4.3
N
H
CH
CH
< 4.4
6.6
990
N
0.3
87
> 300
N
H
*
*
*
*
= SEM within 0.08.
6
* = (µL/min/10 cells).
** = (µL/min/mg protein).
*** = SEM within 0.02.
Given the crystal structures of GSK-3β and CDK-2, as well as the selectivity gained from the
1
presence of a cyclopropyl group at the R position in 10n, we performed a methyl scan around the
pyrimidine ring to assess potency changes against T. b. brucei. The 5 and 6 positions on the pyrimidine ring
were positioned in a similar part of the binding pocket in the GSK-3β crystal structure as the cyclopropyl
group of 10n. Since substitution pointing into the binding pocket of the human enzymes had given
selectivity for 10n, we hoped that substitution off of the pyrimidine ring would give us a similar selectivity
advantage. Using the same chemistry shown in Scheme 1, two analogs were prepared from the starting
pyrimidines with methyl groups at the desired positions. As shown in Table 5, substitution at the 5-position
of the pyrimidine ring (19b), lowered the activity of the series by 10-fold compared to the original hit (1)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
whereas substitution at the 6-position (19a), resulted in a complete loss of activity. In the published crystal
structures and docking models against GSK-3β and CDK-2, these two positions on the pyrimidine ring had
pointed into the same part of the binding pocket as R1.26, 31 While substitution at this position may have
resulted in increased selectivity against the human enzymes, the loss in activity against T. b. brucei indicated
that there was no room for substitution on this portion of the scaffold.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 5. Analogs for methyl scan of pyrimidine.
N
N N
X
N
Y
N
N
H
CN
T. b. brucei Aq. Sol. LogD
RH
Clint
17
80
HLM PPB MRC5
ID
9a
X
Y
*
**
***
pEC50
< 4.4
7.0
(µM)
0.8
0.2
7.4
4
4
Clint
(%) pTC50
1
-H
-CH3
-H
56
170
97
99
< 4.3
< 4.3
19b -CH3
* = SEM within 0.09.
6
** = (µL/min/10 cells).
*** = (µL/min/mg protein).
ADME Optimization
1
2
Armed with the knowledge of the substitutions that were tolerated at the R and R positions, our
3
attention turned to the exploration of the SAR around the R position of our hit compound, 1, with the view
3
to determining which R substituents would provide the most favorable combination of potency and ADME
properties (Table 6). Compound 20a, and 20b demonstrated the importance of the HBD for potency against
T. b. brucei. Removal of the nitrile (10e) did not cause any significant change in the ADME properties of
the series, though replacement with a trifluoromethyl (20c) improved the clearance of the series in both RH
and HLM, though it also resulted in a 4.4-fold decrease in potency against T. b. brucei. Ortho-methyl
ACS Paragon Plus Environment

Page 17 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substitution of the phenyl ring (10g) provided a substantial improvement in the aqueous solubility;
comparison of 10g’s melting point (mp = 177-178 °C) to the melting point of 10e (mp = 207-209 °C)
suggested that this increase in solubility was due to the disruption of the crystal packing of the compound.
However, 10g showed a decrease in the activity against T. b. brucei. To further explore whether ortho-
substitution was a viable route to improve the ADME properties while maintaining potency, we investigated
compounds with a fluorine atom at either the 2- or 6- position of the ring (20d, 20e). These analogs did not
show any significant improvement in the solubility of the series (compared with 1). Next, a number of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
heterocyclic replacements for the R phenyl group were prepared, including pyridines (20f, 20g, 20h),
pyrimidines (20i) and pyridazines (20j). The 3-pyridyl analog (20g) showed a 3.5-fold reduction in the
potency compared to 1 but showed improvements in the aqueous solubility and RH Cl (though HLM Cl
int
int
was still high). As blocking the 4-position of the ring (10s) had previously shown reduced HLM clearance,
0k was prepared, though it demonstrated increased HLM and RH clearance over 20g. The pyrimidine
2
analog (20i) displayed further improvement in solubility and HLM clearance, though this was at a 4.9-fold
cost to potency when compared to 20g.
3
In addition to aryl substituents, a variety of alkyl substituents were explored at the R position, as
3
increasing the sp carbon content has been previously shown to improve the aqueous solubility of a chemical
series.34-35 As the steric bulk and lipophilicity of the compounds increased, an increase in potency against
T. b. brucei was observed, however this increase in potency typically came with a significant reduction in
solubility. Compound 20q was the most active of the compounds against T. b. brucei but demonstrated low
solubility (2 µM) and high intrinsic clearance against both RH and HLM. Introduction of a polar group at
the 4-position was trialed as this region was known to be solvent exposed in human enzymes and we had
3
26, 31
observed a similar SAR relationship between T. b. brucei and these human targets at the R position.
The introduction of a trans-alcohol (20r) greatly improved the ADME profile whilst simultaneously
maintaining the activity against T. b. brucei (pEC : 6.8) as seen with 20q.
50
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
Table 6. T. b. brucei activity and ADME properties of analogs with substituents at the R position.
N
N N
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
R3
T. b. brucei Aq. Sol. LogD
RH
Clint
HLM
Clint
PPB
(%)
MRC5
pTC50
ID
1
R3
*
**
***
****
pEC50
(µM)
7.4
8.2
< 0.1
3.6
66
210
13
98
90
96
< 4.3
N
H
CN
20a
< 4.7
5.9
2
4
2.4
3.0
41
< 4.3
< 4.6
O
N
CN
CN
2
0b
240
200
10e
10g
20c
7.4
5.7
7.6
7.2
0.8
38
3
3.1
3.4
> 4.3
3
62
43
30
50
> 300
250
76
96
96
< 4.3
< 4.3
< 4.3
< 4.3
N
H
N
H
100
nt
N
CF3
CN
H
2
0d
0.4
180
N
H
F
F
2
0e
7.3
6.2
1.4
0.4
3.1
5
57
70
240
8.8
nt
< 4.3
5.7
N
H
CN
10s
100
N
H
2
0f
6.0
7.7
5.3
7.0
7
6
2.4
2.5
-0.6
2
43
5
300
69
3
93
89
41
84
< 4.3
< 4.3
< 4.4
< 4.6
N
H
N
2
0g
N
N
N
H
2
0h
350
26
3.5
35
N
H
N
20i
32
N
N
H
ACS Paragon Plus Environment

Page 19 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
N
N
2
0j
5.1
7.7
840
5
-0.6
2.9
1.3
38
3
44
92
< 4.3
< 4.3
N
H
2
1
0k
N
180
N
H
2
0l
N
OH
5.9
4.8
170
950
1.3
1.3
190
38
25
12
59
40
< 4.3
< 4.3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H
N
H
N
0b
O
O
2
0m
5.4
5.2
1
2
> 3.8
4.1
> 300
94
110
nda
< 4.3
< 4.3
N
H
N
1
0a
> 300
99
H
Cl
20n
20o
4.8
5.3
360
270
1
31
56
34
58
86
< 4.3
< 4.3
NH2
N
H
1.7
220
20p
6.3
6.8
88
2
2.6
3.7
98
> 300
240
85
97
< 4.3
< 4.3
N
H
2
0q
140
N
H
OH
20r
6.8
54
2
8.7
14
68
< 4.3
N
H
*
*
*
*
= SEM within 0.14.
* = (µL/min/10 cells).
** = (µL/min/mg protein).
*** = SEM within 0.17.
6
nt = not tested.
nd = not determined.
low recovery/low stability.
a
Comparison of the initial human kinase and T. b. brucei data indicated that substitution at the R1
position provided a promising opportunity to obtain selectivity. We also wanted to evaluate the impact that
substitution at this position would have on the ADME properties of the series. Compound 21d demonstrated
sub-micromolar activity against T. b. brucei and an improved ADME profile (Table 7), and the
corresponding methyl ether (21e) showed that the presence of the HBD did not have a significant effect on
the potency. The secondary alcohol (21f) and the isopropyl (21g) derivatives were equipotent, though a
slight boost in potency was observed with the cyclopropyl (21h), suggesting that there are steric limitations.
Compound 21h also demonstrated an improvement in the solubility and clearance compared to 1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
Additionally, the cyclopropyl group at the R position was previously shown to cause a total loss of activity
against GSK-3β and CDK-2 (Table 2; 10n).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Table 7. T. b. brucei activity of analogs with a substituent at the R position.
N
N N
R1
N
N
N
H
CN
Aq.
Sol.
T. b. brucei
LogD
7.4
RH
Clint
HLM PPB
MRC5
pTC50
ID
R1
*
**
***
****
pEC50
Clint
(%)
(µM)
2
1a
5.9
nt
nt
nt
nt
nt
< 4.9
MeO
F3C
2
1b
6.2
4.9
0.2
12
4.8
2.8
37
25
16
46
97
95
< 4.3
< 4.3
21c
HO
21d
21e
HO
O
6.5
6.2
20
0.9
2.5
3.3
63
53
240
> 300
93
99
< 4.3
< 4.3
HO
21f
6.1
6.2
6.7
2
2.9
4.4
27
83
18
150
100
80
nt
nt
< 4.3
4.5
2
1g
28
11
2
1h
> 4.2
> 99
< 4.3
*
*
*
*
= SEM within 0.08.
* = (µL/min/10 cells).
** = (µL/min/mg protein).
*** = SEM within 0.03.
6
nt = not tested.
2
Substitution at the R position with a morpholine or methoxy had previously shown a reduction in
activity versus human CDK-2 and CDK-4, as such, we wanted to assess whether these substituents could
ACS Paragon Plus Environment

Page 21 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
also help improve the ADME properties of the series (Table 8). The morpholine was found to have no
effect on the potency of the series, (22c vs. 1) though small improvements in the HLM clearance were
observed. The methoxy group was found to lower potency by 1-log unit (22d vs. 1) against T. b. brucei but
had previously been reported to lead to a total loss in activity against CDK-4 (Table 2; 10j).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Table 8. T. b. brucei activity, and associated ADME profile of analogs with a substituent at the R position.
R2
N
N N
N
R3
N
N
H
T. b. brucei Aq. Sol. Log
RH
HLM PPB MRC5
ID
R2
R3
*
**
***
pEC50
7.5
7.7
(µM)
< 0.9
< 1
D 7.4 Clint
4
4.1
Clint
74
72
100
(%)
98
pTC50
< 4.3
< 4.3
< 4.6
22a
-H
-3-OMe
-3-CN
37
47
66
O
2
2
2b
2c
2d
N
99
a
8.3
< 19
> 3.7
nd
O
2
-3-CN
7.1
nt
nt
nt
nt
nt
< 4.3
*
= SEM within 0.13.
6
** = (µL/min/10 cells).
*** = (µL/min/mg protein).
nd = not determined.
nt = not tested.
a
low MS/MS signal.
Given the observed trends in selectivity, activity, and ADME we set out to make a series of
1
combination analogs (Table 9). For the R position we examined the cyclopropyl group and primary alcohol
as these groups were small enough to not significantly impact the potency of the molecule and could
3
potentially provide a selectivity advantage. We selected three different R groups: 3-aminobenzonitrile, 3-
aminopyridine, and trans-4-aminocyclohexanol based on the observed potency and ADME properties they
displayed.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 9. T. b. brucei activity and ADME properties for a series of combination analogs, combining various
1
2
3
R , R and R substituents.
R2
N
N N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R1
N
R3
N
N
H
T. b.
brucei
pEC50
Aq.
Sol.
(µM)
Log
D 7.4 Clint
RH
HLM
Clint
PPB MRC5
ID
R1
R2
R3
*
*
***
(%)
pTC50
*
O
23a
23b
HO
7.2
0.4
3.6
3.3
nda
3.3
3.4
2.5
3.6
3.8
3.1
2.6
170
45
24
99
< 4.3
CN
CN
CN
O
O
HO
6.8
7.1
6.8
7.2
6.6
7.2
6.8
7.9
6.8
< 0.3
0.4
20
74
nt
100
100
98
< 4.3
< 4.3
< 4.3
< 4.3
< 4.3
< 4.3
< 4.3
< 4.3
< 4.3
N
N
2
3c
3d
3e
23f
3g
3h
49
2
-H
16
15
70
110
45
13
14
60
N
O
2
-H
ndb
13
120
22
97
N
N
N
N
N
O
O
HO
94
2
ndb
2
230
110
53
100
99
O
O
2
N
N
2
3i
-H
0.8
7
94
OH
O
2
3j
-H
-H
7.3
88
N
O
OH
OH
2
3k
6.3
6.3
8.2
2.6
3.4
90
50
52
85
95
< 4.3
< 4.3
O
2
3l
ndb
100
ACS Paragon Plus Environment

Page 23 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OH
O
2
3m HO
5.3
nt
nt
nt
nt
nt
< 4.3
*
*
*
= SEM within 0.13.
6
* = (µL/min/10 cells).
** = (µL/min/mg protein).
nd = not determined.
nt = not tested.
poor MS response.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
no compound detected.
The series of matched analogs detailed in Table 9 gave further insight into the SAR of the cluster.
1
At the R position, analogs with a cyclopropyl group (23c, 23g, 23l) were more potent than the matched
pair analogs containing a primary alcohol (23b, 23f, 23m, respectively). In addition, comparison of 23g
1
and 23l to their matched pair analogs with no substitution at the R position (23e and 23k respectively)
showed no decrease in potency. Interestingly this trend was not observed with compounds containing a
2
morpholine group at the R position, as 23h was over 10-fold less active than its unsubstituted matched pair
1
2
3i. It should also be noted that the unsubstituted analogs at R showed the lowest clearance against both
1
RH and HLM, indicating that R substituents may be a site of metabolism.
2
At the R position, analogs with a methoxy group 23a and 23g were more potent than their
morpholine matched pair analogs 23b and 23h, respectively. The reverse of this trend was seen with 23k
compared to 23j (and 22c vs. 22d previously). This reversal in trends seems to further indicate that
1
2
substitution at the R position could have an impact on the compound-target interactions at the R position.
1
2
Analogs with substitution at R appear to be more potent with a methoxy group at R , while analogs without
1
2
substitution at R were more potent with the morpholine. Compound 23g, with a methoxy group at R ,
showed an increased potency in comparison to the analog with no substitution, 23d. Compound 23d was
equipotent against T. b. brucei with the morpholine derivative, 23h. This increase in potency from the
introduction of a methoxy group can also be observed comparing 21d (Table 7) to 23a, where 23a was 4-
fold more active. Overall the morpholine group was better for controlling the clearance of the series with
matched analogs (23h and 23j) containing the morpholine showing lower RH and HLM clearance than the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
methoxy counterparts (23g and 23k), though all R analogs showed higher clearance and were less soluble
than the unsubstituted analogs (23d and 20r).
A subset of these combination molecules was sent for single point inhibition assays against all three
human kinases (Table 10). Analogs were prioritized based on their predicted selectivity. As expected, 23i
and 23e displayed the highest percent inhibition of all three human enzymes, reaffirming the need for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
substitution at R to provide selectivity. This was also evident when examining the selectivity trend for
1
1
CDK-4, as any substitution at R caused a significant decrease in the activity against the enzyme. At R the
primary alcohol (23a, 23b) was generally less active than the cyclopropyl (23d, 23g, 23h) against CDK-2.
1
The morpholine group did appear to decrease affinity for CDK-2 provided there was substitution at R (23h,
1
23b). For GSK-3β the cyclopropyl group at R provided the greatest decrease in percent inhibition (23d,
2
23h), however the morpholine at R seemed to increase potency against the human enzyme. The most
selective of these compounds, 23a, displayed less than 50% inhibition against all three targets, however
due to a poor ADME profile it was not considered for further evaluation.
Table 10. Selected analogs % inhibition at 1 µM against human kinases CDK-2, CDK-4, and GSK-3β.
CDK-2
cyclin A cyclin E
CDK-2
CDK-4
cyclin
D3 (%)
1
T. b. brucei
GSK-
3β (%)
Cpd
3a
*
pEC50
(
%)
47
46
76
84
84
67
89
(%)
29
54
70
84
77
49
87
2
7.2
6.8
6.8
7.2
7.2
6.8
7.9
49
75
21
89
65
34
81
23b
23d
23e
32
21
69
25
26
90
2
2
2
3g
3h
3i
*
= SEM within 0.13.
With evidence of what was required to obtain selectivity in the series, we set out to determine
whether the pyrazolo[1,5-b]pyridazine head would be efficacious in an animal model of infection. Based
ACS Paragon Plus Environment

Page 25 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
on the observed SAR and ADME data, summarized in Figure 3, we prioritized 20r, and 20g, for further
investigation. Further studies showed that both 20r and 20g were stable in mouse plasma (Table S3,
Supplemental Information), were not substrates of the cytochrome P450 isozyme 3A4, and did not induce
testosterone metabolism in CYP450 3A4 (Table S4, Supplemental Information). Both compounds were
also found to be permeable in Caco-2 cells and were not actively effluxed (Table S5, Supplemental
Information).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A small number of compounds were assessed to see if they were still fast acting and parasiticidal,
including 20r and 20g (Table S6, Supplemental Information). Using a set of criteria previously
1
8
disclosed, inhibitory properties of both compounds were assessed at set time points (Figure 4); we defined
fast-acting compounds as having a pEC ≥ 6 after 18 h of incubation. Given this definition, 20g was fast-
5
0
acting and 20r was slow acting. In parallel, washout assays were performed to assess whether the
compounds were cidal or static. The compounds were incubated for 18 h with T. b. brucei then washed out
and the parasite was incubated for 72 h before the wells were assayed to determine pEC values. Compound
5
0
2
0g was deemed cidal as it demonstrated a pEC ≥ 6 following the washout assay, whereas 20r was not
99
cidal.
Figure 3. Summary of SAR and SPR relationships for the pyrazolo[1,5-b]pyridazines.
R1
R2
- Methoxy is more active if substitution at R1
- Morpholine more active if no substitution at R1
-
-
Primary alcohol gives best
overall selectivity
-
No substitution gives
poor selectivity
O
N
N N
R3
Large groups not tolerated
-
Aliphatic groups displayed better
ADME properties
HO
N
-
Aromatic more active than aliphatic
EWG at 3-position are more active
- No selectivity improvement
N
N
CN
-
Pyrimidine and NH
Critical for T. b. brucei activity
H
-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 4. Rate of action of 20g and 20r, comparing pEC values against T. b. brucei following defined
5
0
periods of incubation with the compound. Compounds with a pEC ≥ 6 at 18 h were deemed fast acting
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(shown with a darkened line).
8
7
.5
7
6.5
6
2
0g
20r
5
.5
5
6
h
12h
18h
24h
72h
Incubation Time
Both 20g and 20r were advanced to pharmacokinetic (PK) studies in mice, 20r for its superior
ADME profile, and 20g for its rate of action, cidal activity, and high potency. Although neither compound
contained structural features that were indicative of selectivity over human GSK-3β, CDK-2 and CDK-4,
we theorized that their ADME profiles and activity gave the highest probability of providing proof of
concept efficacy. With the knowledge that inhibitors of GSK-3β have been shown to have no negative
effect in in vivo studies, and the lack of activity observed against human MRC5 cells for either 20g and
2
0r, we believed that there may be an acceptable therapeutic window.19, 36
Following a 10 mg/kg intraperitoneal dose (IP) into female NMRI mice, 20r (Figure S2,
Supplemental Information) showed low whole blood exposure, with levels above the EC value (152 nM
5
0
ACS Paragon Plus Environment

Page 27 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
or 47 ng/mL) for < 4 h following IP administration. In addition, the compound levels were below the lower
limit of quantitation in the CNS after 4 h. We therefore deprioritized 20r for efficacy studies.
Compound 20g demonstrated much better BBB penetration with similar concentrations in both the
blood and brain (Figure 5). In addition, although the compound displayed a short half-life (t = 0.68 h),
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
/2
blood and brain concentrations were > 10×EC value (20 nM or 6.1 ng/mL) for the entire duration of the
5
0
study. Given the window for activity, as well as BBB penetration 20g was taken forward into an efficacy
study.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 92
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 5. Peripheral blood levels of 20g after IP administration to female NMRI mice (n=3) at a target
dose of 10 mg/kg in 1% DMSO, 20% Captisol® in water. Individual values for each time point are
represented in the plot.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2500
Blood Concentration (ng/mL)
2000
1500
1000
Blood Concentration (ng/mL)
Brain Concentration (ng/g)
500
0
0
1
2
3
4
5
Time (hr)
For the efficacy study, mice were infected with T. b. brucei STIB795 and three days post-infection
0g was administered at 10 mg/kg/day IP for five consecutive days with daily monitoring of parasitemia
2
levels. A significant reduction was observed in parasitemia for four of six mice, resulting in parasite
concentrations dropping below detectable levels by day 4 of treatment (Table S10, Supplemental
Information). Despite the reduction of parasitemia, 20g showed a negligible improvement in the survival
rate of the mice compared with the control group, likely due to the toxicity of the compound, as mice
presented with conjunctivitis, lacrimation, facial inflammation, in addition to an inflamed liver and kidneys
(Figure 6).
ACS Paragon Plus Environment

Page 29 of 92
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 6. Effect of 20g on trypanosomes in a mouse efficacy model of HAT. Percent (%) survival rate of
mice dosed IP at 10 mg/kg/day, compared to the control vehicle.
1
00
90
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
80
70
60
50
40
30
20
10
0
Control (Vehicle)
0g - 10 mg/kg/day
2
1
2
3
4 Days 5
6
7
8
Due to the toxicity observed from the efficacy study we deemed it pertinent to examine the wider
kinase selectivity of 20g, profiling it against a panel of human kinases (Figure 7). Of the 47 kinases 20g
was profiled against, fourteen showed greater than 50% inhibition at 1 µM. Unsurprisingly, given the lack
1
2
of substitution around R and R , 20g still showed potent inhibition of GSK-3β, CDK-2 and CDK-4. Further
efforts in the optimization of this series would need to address this toxicity. Given the improvement in the
1
2
selectivity of 23a a more thorough investigation of substituents at R and R is warranted. The difference
in toxicity, shown in Table S11, Supplemental Information, against L6 (20g pTC > 6.2, 23a pTC <
5
0
50
4
.3) and THP-1 (20g pTC > 6.2, 23a pTC < 4.3) indicate that these may be better cell lines to use when
50 50
optimizing for toxicity.
Figure 7. Percent (%) inhibition at 1 µM of 20g against human kinases. Line drawn to show 50% inhibition
of enzymes. Kinases showing significant inhibition (above 50%) are listed on the X-axis. We define
ACS Paragon Plus Environment